gg (@_gianfg) header
gg (@_gianfg) profile picture

the book of nature is written in the language of mathematics

Argentina
Joined January 2012
Photo by gg (@_gianfg)Photo by gg (@_gianfg)Photo by gg (@_gianfg)Photo by gg (@_gianfg)Photo by gg (@_gianfg)Photo by gg (@_gianfg)Photo by gg (@_gianfg)Photo by gg (@_gianfg)Photo by gg (@_gianfg)Photo by gg (@_gianfg)
Tenemos la misma foto de perfil
Si calculas 1/2 kg por persona en un asado entre hombres, tambien calculense la medida de la cintura asi les entra la minifalda o la pollerita
1
1
1
85
Free Claude or Codex credits to do research with AI agents, plus a trip to present it in Paris. It's called the Lightcone Research Fellowship, from UC Berkeley and CNRS, backed by Eric and Wendy Schmidt. What you get: • Free Claude or Codex model credits for your research • Direct Slack access to their scientists and engineers • A tiny cohort of 4–6 fellows from different fields • An invite to present at the Paris Agentic AI for Science Summit, with travel support available • Fellowship credit for your CV What you don't need: • To be in a specific field. Every discipline is welcome • A brand new project. One already underway, or almost done, works too • To move anywhere. It's fully remote The condition: you're an early-career researcher with a clear project you want to run with AI agents and your supervisor signs off. Deadline: Nov 20, 2026. Apply here: lightconeresearch.org/fellow… Bookmark it.
7
17
1
134
8,510
after a year of work, I am here presenting a new university degree we have been developing at the University of Buenos Aires a degree in Complex Systems, designed from scratch for a world where AI is radically transforming how we work, learn, and create wish us luck!
72
35
13
546
35,320
Lula es lo que los peronistas piensan que Perón quería ser
17
Yo no seré un grandmaster pero ese opening que le recomendo es una verga
Messi asked ChatGPT to teach him a chess opening that matched his style. It suggested Goat Horns 🤘
20
Learn to build evals, folks. Highly recommended reading on using decision models for evals. From our internal evals, System One models like Jev are a great fit for LLM Judges. If you want to go beyond Hamel's great write-up, I've also put together a few resources on applying decision models to LLM judges. (bookmark it) 1) Jev-as-a-Judge for Agent Evaluations - academy.dair.ai/dashboard/re… 2) Jev-as-a-Judge: Evaluating Agents with a Decision Model - academy.dair.ai/labs/jev-as-… Bonus paper I would recommend, with a few other ideas that combine System One and System Two models: Jev-as-a-Judge paper - arxiv.org/abs/2609.26550
New blog post 🎉: How To Use Decision Models For Evals In this post, we show you how to choose a decision model for your task, tune its threshold (don't assume the 0.5 default is good!), and interpret its scores. Lots of cool interactive charts here.
69
151
7
1,666
128,742
Es increíble como tomar un poquito de sol te cambia el día
14
La gente a la hora vender productos Apple se pone muy retrasada. No podes vender una MacBook Air 2014 a 400 dolares.
15
现在顶尖大学里教 Agent,已经开始一路教到 GPU 了。 非常推荐 Harvard 的 CS2680 Modern AI Systems 。课程从 Claude Code、Agent Loop、Eval 和 Token 成本开始,一路往下做到模型部署、Routing、KV Cache、量化和 speculative decoding。 作业设计也很有巧思,5 个作业都围着同一个 Agent 系统往下拆,先看它怎么工作,再自己搭 Harness,最后一路优化到推理层。 感觉它更像是在回答一个问题: 一个现代 AI Agent,到底是怎么从应用层一路跑到底层系统的? 做 Agent、Harness、AI Infra 的可以收藏。 cs2680.seas.harvard.edu/
120
930
42
5,492
386,001
This semester I am teaching a new course on Agentic Systems, which explores the full stack of AI systems: from building and optimizing agent harnesses to serving and optimizing LLM serving infrastructure. Students learn by building, measuring, and optimizing real agents and systems, balancing capability, latency, and cost. Our goal is not just to teach today's AI technologies, but to prepare students to build tomorrow's. All lectures, readings, and assignments are publicly available at cs2680.seas.harvard.edu
70
448
15
4,007
168,189
El plan de $20 de claude es la versión de demo
11
reading feynman lecturess as feynman intended: with accompanying ai-made demo sections for each chapter rapid acquisition of knowledge is possible thanks to superintelligence
97
476
48
4,047
195,281
Me prestan 5k?
a full humanoid robot used to cost more than a car. berkeley just made one for under $5k. berkeley humanoid lite changes everything: > modular, 3d-printed gearboxes > off-the-shelf hobby parts > cad, firmware, and the full rl training stack included a five-figure parts order becomes a weekend of printing. that’s the difference between reading about humanoid robots and having one on your bench. 100% open-source.
10
Our Stanford class "CME 295: Transformers & Large Language Models" is back for a new edition and Lecture 1 is now on YouTube!
3
54
5
502
16,483
Stanford dropped CS329Z today. Engineering AI Agents. compound systems, build RAG/tool use from scratch, then frameworks, then how you actually evaluate this stuff. gonna follow along online. looks pretty neat. cs329z.stanford.edu/
13
185
3
1,475
87,417
We're flying students to the biggest blockchain student conference on Earth. We partner with @UBC_Conference to fly LATAM students to the conference. 1,000+ students from 100+ universities will be there. In two years, UBC has sent 80+ students into frontier tech, with teams joining Y Combinator, a16z Speedrun, and Colosseum. This opportunity could change your life, apply now. becas-ubc.vercel.app/
7
20
6
70
6,428
This
My Stanford course 𝗧𝗵𝗲 𝗠𝗼𝗱𝗲𝗿𝗻 𝗦𝗼𝗳𝘁𝘄𝗮𝗿𝗲 𝗗𝗲𝘃𝗲𝗹𝗼𝗽𝗲𝗿 starts next Tuesday. This Github repository will hold all the assignments (and has the ones from last year). Bookmark it. See you next week. github.com/mihail911/modern-…
19
El futuro es completamente open source
The robotics company’s prisoner’s dilemma: don’t use GPT Astra and your competitors will show better demos and capabilities. Use it, and your data & moat will be absorbed by the next version of the LLM.
23
We are hiring at Supabase! There are 60 open positions across engineering, sales, design, and many more, so there should be something that would fit you! We are a remote company, so where you live doesn't matter. If you are awesome at what you do, we'd love to work with you!
127
96
11
1,958
155,652
próximamente open-source
🚨Revolut confirmó que entregó datos sensibles de algunos clientes tras una solicitud gubernamental falsa: según CoinBureau, los atacantes exigen 10.000 bitcoin (cerca de US$780 millones) para frenar nuevas filtraciones. 👉🏻La empresa dice que los fondos están intactos ⚠️Qué datos se expusieron 🔹 Nombre, fecha de nacimiento, domicilio, mail y teléfono 🔹 Copias de pasaporte o licencia y selfies de verificación 🔹 IBAN, extractos, retiros e historial de movimientos, incluido bitcoin
10